The long-term goal of this project is to develop a robust and hardened high-resolution air-ocean coupled tropical cyclone (TC) data assimilation and prediction system that is able to assimilate the wide variety of available in-situ and remotely-sensed observations in order to analyze and predict TC structure and intensity changes in an operational environment. Significant gains have been made in TC track prediction over the past three decades. This considerable achievement is due, in large part, to the steady improvement of numerical models, especially the global scale prediction systems, and the judicial utilization of multi-model ensemble results. In contrast, the TC intensity forecast by numerical models has shown very little improvement during the same time period, and remains a formidable forecast problem. Advanced statistical prediction models nowadays are able to predict the trend for intensification, but as statistical tools, they inherently cannot predict the rapid intensity changes, as evident in Katrina and Rita of 2005, and other tropical cyclones. It is generally accepted now that while advancements in data assimilation and modeling have resulted in better analyses and predictions of steering flows, the processes that affect the structure and intensity of tropical cyclones are much more difficult for current numerical models to capture and reproduce. Physical processes in tropical cyclones that can affect their structure and intensity include enthalpy and mechanical interchanges with the underlying ocean and land surfaces, shallow and deep atmospheric convection in the convectively unstable tropical atmosphere with vertical and horizontal wind shears, and internal multiscale non-linear dynamic interactions. Current prediction systems have been shown to be able to reproduce rapid intensification in case studies involving complex upper tropospheric and oceanic conditions in a carefully conducted simulation mode (e.g. Hong et al. 2000).